Journal of Agricultural Engineering 2013; volume XLIV:e16
A low-cost electronic tensiometer system for continuous monitoring
of soil water potential
Martin Thalheimer
Research Centre for Agriculture and Forestry Laimburg, Bolzano, Italy
Abstract
A low cost system for measuring soil water potential and data log-
ging was developed on the basis of an Arduino microcontroller board,
electronic pressure transducers and water-filled tensiometers. The
assembly of this system requires only minimal soldering, limited to the
wiring of the power supply and the pressure sensors to the microcon-
troller board. The system presented here is, therefore, not only inex-
pensive, but also suited for easy reproduction by users with only basic
technical skills. The utility and reliability of the system was tested in a
commercial apple orchard.
Introduction
Increasingly, problems related to water resources are the cause of
us agricultural and horticultural crops. Electronic devices have made it
al
limitations in agriculture. Therefore, the availability of simple yet
accurate tools for monitoring plant-available water and scheduling irri-
ci
gation will become a key factor for saving significant amounts of water
er
and energy while maintaining, or even improving, yield and quality of
agricultural produce in many climatic regions of the world.
m
For many decades, this topic has been the subject of intensive
research, and a wide range of methods for monitoring soil moisture
m
has been proposed (Evett, 2008). These include techniques to measure
soil water content or potential as well as soil water budgets calculated
co
on the basis of soil, plant, and climatic data.
Tensiometers are the only instruments capable of providing a direct
-
measurement of soil matric potential due to the hydraulic connection
on
N
Correspondence: Martin Thalheimer, Research Center for Agriculture and
Forestry Laimburg, Laimburg 6, 39040 Auer/Ora (BZ), Italy.
Tel: +39.0471.969652.
E-mail: martin.thalheimer@provinz.bz.it
Key words: soil water potential, microcontroller, pressure transducer, ten-
siometer, data logging, Arduino.
Received for publication: 28 August 2013.
Accepted for publication: 6 November 2013.
©Copyright M. Thalheimer, 2013
Licensee PAGEPress, Italy
Journal of Agricultural Engineering 2013; XLIV:e16
doi:10.4081/jae.2013.e16
This article is distributed under the terms of the Creative Commons
Attribution Noncommercial License (by-nc 3.0) which permits any noncom-
mercial use, distribution, and reproduction in any medium, provided the orig-
inal author(s) and source are credited.
[page 114] [Journal of Agricultural Engineering 2013; XLIV:e16]
between soil water and the water inside the tensiometer shaft. Soil
water potential is thus translated into a negative pressure inside the
tensiometer that can be sensed by a mechanical gauge or electronic
pressure sensor. At equilibrium, the partial vacuum inside the ten-
siometer equals the sum of soil matric potential plus the usually neg-
ligible effect of the height of the tensiometer’s water column. Because
of the direct relation of matric water potential to plant water availabil-
ity, tensiometers require no site-specific calibration and are, there-
fore, well suited for a wide range of applications in soil hydrology and
y
irrigation management. A frequently mentioned inconvenience of con-
l
ventional water-filled tensiometers is their limited range of operation,
on
typically within 0 to −80 kPa. However, most crops present their soil
moisture optimum well within this limit (Gratton and Oster, 1992;
Sanders, 1997; Thompson et al., 2007). Tensiometers are, therefore, an
e
excellent tool for optimizing water-use efficiency of a multitude of
possible to implement systems of automatic data acquisition and irri-
gation control. Commercially available systems are, however, often
expensive and sometimes complicated to operate. Recent advances in
electronics have led to the development of inexpensive components
which can be assembled into fully operational sensing, monitoring,
and control systems (Dedrick et al., 2000; Fisher and Gould, 2012;
Fisher and Kebede, 2010; Greswell et al., 2009; Riley et al., 2006).
The main objective of this paper is to propose an easily reproducible,
low cost tensiometer system that requires minimal soldering of elec-
tronic components. It describes the technical layout and the assembly
of the electronic unit as well as the assembly of the tensiometers from
easily available materials. The presented system is based on the open-
source microcontroller platform, Arduino, and a few additional compo-
nents, and it is designed to accomplish accurate measurements of soil
matric potential at regular intervals and to store data in memory for
later retrieval. A suitable software code for the microcontroller is freely
available. The system has been successfully tested for research pur-
poses and in commercial fruit production.
While combinations of data loggers and electronic tensiometers
have been commercially available for many years, their cost has
restricted their use mainly to research purposes or high value crops.
The very limited cost of the system presented in this paper could rep-
resent a step towards a more widespread application of accurate mon-
itoring of soil water potential, leading to an improvement of water use
efficiency in many fields of agricultural production.
Materials and methods
The core of the data acquisition and logging system presented here
consists of a microcontroller board and a variable number of electronic
pressure transducers. Power for prolonged standalone use is provided
by a small solar panel and a re-chargeable lead acid battery. The hard-
ware components are shown in Figure 1 and the schematic diagram of
the circuit is presented in Figure 2.
 Technical Note

Microcontroller board where:

hw is the height of the hanging water column between the sensor level
There is a wide choice of commercially available microcontrollers.
and the water meniscus in the tube;
Most of them require additional external components and/or specific
g the gravitational force;
ρw the density of water.
programming interfaces to become fully operational units, and there-
fore demand a substantial degree of technical expertise both for assem-
The resulting calibration function (Figure 3) describes a perfect lin-
bly and programming. In recent years, an open source project has been
earity between pressure differential and sensor output (r2=1). The
launched, designed to make the use of microcontrollers in multidisci-
experimentally established sensitivity of the sensor was of 44 mV/kPa,
plinary projects more easily accessible. The Arduino system consists of
thus coming very close to the sensitivity of 45 mV/kPa stated in the offi-
boards equipped with Atmel AVR microcontrollers and complementary
cial technical datasheet.
components needed for regulated power supply, serial communication
and interfacing with external hardware. The microcontroller is
equipped with a bootloader, which enables programming via open-
source PC software without additional specific hardware.
The presented technical layout is based on an Arduino board with an
Atmega328 microcontroller and 5V power supply. The Atmega328 fea-
tures 6 analogue (A/D) ports (10 bit resolution) with an input range of
0-5 V, 14 digital input/output ports, and 32 KB of flash memory, as well
as 1 KB EEPROM. Each A/D port of the microcontroller can be used as
input for an electronic pressure sensor, thus allowing a maximum of 6

y
tensiometers per unit.
As with the hardware, the Arduino integrated development environ-

l
on
ment is open source and can be down-loaded freely from the project’s
internet website (http://www.arduino.cc). This application is written in
Java and, besides providing the environment for programming, serves

e
for compiling and up-loading programs to the microcontroller.
Programmes are written in a language similar to C/C++, with some
simplifications and modifications. Detailed reference about boards and
the programming language, including explanatory examples, can be
us
al
found on the Arduino website (http://www.arduino.cc).
ci

Electronic pressure transducers

er

The type of sensor used for this application is the piezoresistive dif-
ferential pressure transducer MPX5100DP from Freescale
m

Semiconductor with integrated temperature compensation and signal

amplification circuits, providing a linear voltage output for a differen-
m

tial pressure range from 0 to 100 kPa. The sensitivity of the sensor cor-
responds to 45 mV/kPa, according to the official product datasheet
co

(Freescale Semiconductor, 2010).

The MPX5100DP sensor requires a power supply of 5 V and provides
Figure 1. Tensiometer station hardware: (A) photovoltaic panel,
-

an output voltage with a full-scale span of 4.5 V and a maximum output (B) re-chargeable battery, (C) Arduino board, (D) assembled (left)
on

of 4.8 V. MPX5100DP sensors have been used for various scientific ten- and unassembled (right) tensiometer, (E) pressure transducers
siometer applications for more than a decade (Testezlaf et al., 1999; wired to the controller board.
Thalheimer, 2003; Grashey-Jansen, 2012).
N

In order to verify the official data about the sensitivity of the sensor,
a detailed output analysis over a differential pressure range of from 0
to 650 hPa was performed on a randomly chosen MPX5100DP sensor.
The output curve in Figure 3 was established with the hanging water
column technique, measuring the sensor output at successive increas-
es in the height of a hanging water column, thus creating an increas-
ingly negative pressure at the sensor level.
For this purpose, several metres length of a water filled, transparent
3x5 mm flexible PVC tubing were attached to the vacuum port of the
sensor and arranged into a U-shape. The free end of the tube was shift-
ed vertically downwards in successive steps such as to move the water
meniscus in the tube from its initial position at the same height of the
sensor to increasingly lower positions. Sensor readings were taken at
each step of 10 cm until a final height of the hanging water column of
650 cm was reached. The pressure difference (ΔP) relative to atmos-
pheric pressure at each step of the procedure was calculated according
to the following formula:

ΔP = hwgρw (1) Figure 2. Electrical scheme of the tensiometer logging system.

[Journal of Agricultural Engineering 2013; XLIV:e16] [page 115]

Technical Note

Power supply
The power consumption of different Arduino and compatible boards
varies according to their specific technical design. Standard Arduino
boards are generally equipped with a number of components not strictly
necessary for standalone operation, such as LEDs and communication
interfaces, resulting in a typical power consumption of around 25 mA.
Arduino boards with reduced external circuitry, e.g. without the USB
interface, have a slightly lower power consumption of around 15 mA.
The MPX5100DP pressure transducers require a supply current of
typically 7 mA (max. 10 mA). In the present system, the pressure sen-
sors are each powered individually by a digital port of the microcon-
troller, thus restricting the power supply to the pressure sensors only to
the short time necessary for each tensiometer reading to be taken.
Further substantial reductions in power consumption are technically
possible, such as implementing a sleep-mode of the microcontroller in
conjunction with an external real-time clock, as proposed in various
similar projects (Afanasyev, 2004; Fisher, 2007; Fisher and Gould, 2012;
Fisher and Kebede, 2010). However, in accordance with the outlined
objective of maximum simplicity and ease of replication, preference
diameter and length. A 16-mm outer diameter PVC pipe of desired
length was fixed with hot-melt adhesive inside the plastic neck of a
porous ceramic cup. The threaded end of the plastic neck had been cut
off previously. Optionally, a piece of 25-mm outer diameter PVC pipe
can be added externally in order to obtain a tensiometer of the same
external diameter as that of the ceramic cup (Figure 4A).
The air entry point of the abovementioned ceramic cups was deter-
mined by direct observation on a representative number of specimens
and was found to be typically within the range of 650 to 850 hPa. The
cups present a hydraulic conductivity that is sufficiently high to allow
rapid response to soil wetting as shown in Figure 5 by the sudden
increases of soil matric potential after events of consistent rainfall.
The electronic pressure sensors were fitted onto silicone stoppers
that serve to open/close the tensiometers for water filling. The silicone
stoppers of appropriate size were perforated and a piece of 3x5 mm
flexible PVC tubing inserted and trimmed to a length of approximately
10 mm protruding at the upper side of the stopper. Finally, the electron-
ic pressure transducer was inserted with its vacuum nipple into the
upper end of this tubing (Figure 4B).

y
was given to a combination of a small photovoltaic panel (5W/12V) and Software code

l
a re-chargeable battery of appropriate size (6Ah) to provide sufficient

on
A software code for this tensiometer station is freely available direct-
power for continuous and long-term operation of the unit. The solar
ly from the Author on request. The code consists of a simple clock and
panel can be directly connected to the battery without any further
calendar schedule based on the microcontroller’s internal time counter.
charge control circuit, if the output current of the panel remains within

e
The programme carries out tensiometer readings at regular time inter-
10% of the battery’s storage capacity. If not already provided on the vals, which can be set to different values (1-2-3-4-6-8-12-24 h) by
panel, only a blocking diode needs to be added in order to prevent
reverse flow of current from the battery through the solar module dur-
ing darkness (Figure 2).
us accessing the user menu provided in the software. An average value is
obtained by calculating the arithmetic mean of 10 individual measure-
ments of each tensiometer, replicated at short intervals. To optimize
al
Data storage and transfer
ci

The Atmega328 is equipped with a 1 Kbyte EEPROM that can be used

er

for data storage. This may not seem very much compared to the storage
capacities of other modern electronic devices, but can provide suffi-
Table 1. Pin connection between MPX5100DP sensors and the
m

cient memory for many practical applications. In the case of a configu- Arduino board. Port numbers in brackets correspond to the
ration with 3 tensiometers and an hourly interval of data-logging, the 1 wiring example shown in Figure 2.
m

Kbyte EEPROM provides data storage for approximately two weeks.

MPX5100DP Arduino
Setting longer intervals between consecutive readings will increase the
co

logging autonomy accordingly. Data storage capacity could also be 1. Signal (notched) AD port (0)
increased by adding an external EEPROM, but this would again involve 2. GND GND
-

a more complex technical design and, therefore, go beyond the scope of

on

3. Power supply Digital port (2)

this project.
Data can be downloaded from the microcontroller board to the PC via
N

a serial connection. Most types of Arduino boards are already equipped

with the required electronic interface. Simpler boards without serial
interface need an external adapter for converting the microcontroller’s
TTL levels to signals compatible with the PC serial port. On the PC, a
suitable terminal programme (e.g. Hyper Terminal) makes it possible
to retrieve and save the data transferred from the microcontroller’s
memory.

Assembly of the electronic unit

The assembly of the electronic hardware requires only the wiring of
the power supply and the electronic pressure sensors to the Arduino
board. On the sensor side, a 3-conductor cable is soldered to pins 1-3 of
each pressure transducer. The other ends of the cable are connected to
the appropriate pins of the Arduino board, as listed in Table 1.

Tensiometer assembly
Tensiometers were assembled from commercially available porous
ceramic cups (Tropf-Blumat, Weninger Kunststoff - Keramikwerk,
Austria; http://www.blumat.at) and transparent PVC pipes of suitable
Figure 3. Experimental calibration curve of an MPX5100DP sen-
sor. The slope of the linear regression of sensor output versus
pressure (sensor sensitivity) is 44 mV/kPa (r2=1).

[page 116] [Journal of Agricultural Engineering 2013; XLIV:e16]

 Technical Note

memory use, the A/D readings of the pressure transducers are reduced
from their original 10 bit to 8 bit resolution before storage. The data Results
transfer function of the programme retrieves the data from the memo-
ry, transforms the original A/D values to pressure units (hPa) on the Apple orchard case study
basis of the typical calibration function provided by the manufacturer A tensiometer station of the described design on the basis of an
of the pressure transducer (Freescale Semiconductor, 2010) and finally Arduino Uno board, and with the logging interval set to 1 h, was
sends them as absolute values to the serial port. The final resolution of installed during the growing season 2012 in an apple orchard at the
the tensiometer data retrieved from the logger is 4-5 hPa, which is eas- Research Centre for Agriculture and Forestry Laimburg in Northern
ily sufficient for most purposes of agricultural water management. Italy (Figures 6 and 7). Three tensiometers were placed along a tree
For setting time, date and logging interval, as well as for retrieving
or erasing data from the memory, a simple menu can be accessed via a
serial connection with a PC. This menu also offers the options of read-
ing actual tensiometer values and the current software settings at any
given time. Furthermore, the menu offers the option of setting a
threshold value for the tensiometer readings of soil water potential. At
any periodic tensiometer reading, the measured values are compared
to the chosen threshold and if, in terms of absolute values, soil water
tension reaches or passes the threshold, the LED connected to digital
pin 13, which is present on most Arduino boards, starts blinking, thus
giving a visual indication e.g. for the need of irrigation. Vice versa,

y
blinking will stop again as soon as tensiometer readings return below

l
the set threshold.

on
The freely available software code contains further operational
details in the form of comments.

e
us Figure 5. Soil water potential recorded by 3 tensiometers at a
depth of 25 cm and daily rainfall for a 6-week period during the
growing season 2012 in a high-density apple orchard.
al
ci
er
m
m
- co
on
N

Figure 4. Schematic presentation of tensiometer makeup: (A) ten-

siometer shaft consisting of a 25 mm outer and 16 mm inner Figure 6. An electronic tensiometer station installed in an apple
PVC pipe, both connected to a ceramic cup; (B) silicone stopper orchard. (A) Photovoltaic panel. (B and C) Microcontroller board
with inserted pressure transducer. Relative size of the two objects and battery, both under protective cover. The tensiometers are
not in scale. hidden by the trees.

[Journal of Agricultural Engineering 2013; XLIV:e16] [page 117]

Technical Note

row at a distance of approximately 2 m from each other and with their
ceramic cups at a depth of 25 cm, which corresponds to the area of
highest root density of intensive apple orchards on dwarfing rootstocks.
The Arduino board and the battery were mounted on a wooden support,
which in its turn was fastened to a cement pole. The photovoltaic panel
was mounted on the same pole, overlooking the tree canopies and fac-
ing south for optimal interception of solar irradiation. A plastic contain-
er of suitable size turned upside down over the Arduino board provided
adequate protection against rainfall. Also, the battery was protected by
a plastic cover.
Data were regularly retrieved to a mobile PC at intervals of no more
than two weeks so as not to exceed the available data storage capacity.
The pattern of soil water potential recorded by the tensiometer station
during a 6-week period between the beginning of August and mid-
September 2012 is shown in Figure 5. The graph reveals the general
pattern of decreasing soil water potential as a result of plant water
uptake over time and sudden increases after rainfall events. The graph
also indicates a marked pattern of daily fluctuations of tensiometer
readings, which result from the diurnal temperature changes deter-
mining the contraction or expansion of the small air volume trapped in
both in the well-distributed rainfall during the observation period and
in the high water table which by capillary rise significantly contributes
to satisfying the plants’ water requirements at this site (Thalheimer,
2005). No irrigation was applied during this period of observation. This
puts into clear evidence how the low-cost monitoring system presented
here can effectively help to avoid unnecessary irrigation cycles or, in
different circumstances, indicate the precise moment when irrigation
is needed.
Conclusions
The proposed electronic data acquisition system for soil matric
potential based on the open source Arduino system, MPX5100DP differ-
ential pressure sensors and self-assembled tensiometers, constitutes a
very cost effective alternative to commercial soil moisture monitoring
equipment. The prices of electronic components are in continuous evo-
lution and can vary enormously depending on the source of supply and
the order volume. An estimate of the cost of the components of the pre-

y
the upper part of the tensiometer shafts (Warrick et al., 1998). sented system is given in Table 2. For example, in the case of a config-

l
During this period of observation, the soil water potential never

reached levels conducive to plant water stress. The reason for this lays
on
uration of the data acquisition system with three tensiometers, as used
in the orchard test described, the estimate of the total cost of the sys-
tem amounts to 115 euros.
From a technical point of view, the system has proved to be accurate

e
and reliable, running for a test period of several months in two consec-
us utive years without any malfunction. As for most electronic equipment,
however, it cannot be excluded that under particular circumstances,
such as strong electromagnetic fields generated by e.g. lightning dis-
al
charge, the stability of the system may be compromised. Periodical con-
trols of the units installed in the open field are, therefore, advisable.
ci

From an agronomic viewpoint, the field testing in an apple orchard

er

confirmed that the presented system has a high potential for substan-
tially improving the efficiency of agricultural water use. This is made
m

possible by the continuous monitoring of the availability of soil water

for plant uptake and by the triggering of a visual signal indicating the
m

precise moment when the need for irrigation arises. Furthermore, the
data logging function allows the soil moisture pattern to be document-
co

ed for prolonged periods and to carry out subsequent analyses of soil

water dynamics as influenced by soil, climate, crop and irrigation.
Figure 7. Detail of the tensiometer station shown in Figure 5, dis-
-

The assembly of a tensiometer station of the described design

on

playing the Arduino Uno board and battery with (left) and the
Arduino board without (right) protective cover. The cables on the
left lead to 3 tensiometers; the black cable on the right side con-
N
requires only minimal soldering and can, therefore, be easily set up
also by users with only basic technical skills. The simplicity of construc-
tion, combined with its modest cost, could make this system a useful

nects to the 12V battery.

tool for studies of soil hydrology in general and for irrigation manage-
ment in particular.

Table 2. Indicative prices (without VAT and shipping) of the com-

ponents of the electronic tensiometer system.
Part Price (€) References
Microcontroller board and power supply
Arduino Uno board 23.5 Afanasyev V. 2004. A miniature daylight level and activity data recorder
12V/6Ah sealed lead acid battery 13.5 for tracking animals over long periods. Mem. Natl. Inst. Polar Res.
5W/12V solar panel 15 58:227-33.
Cable 2 Dedrick R.R., Halfman J.D., McKinney D.B. 2000. An inexpensive,
Total 52 microprocessor-based, data logging system. Comput. Geosci.
Tensiometer 26:1059-66.
MPX5100DP pressure sensor 13.6 Evett S. 2008. Field estimation of soil water content: a practical guide
PVC pipes for a tensiometer of 50 cm length 2.6 to methods, instrumentation and sensors technology. Training
Tropf-Blumat cone 2.3 Course Series No. 30, International Atomic Energy Agency, Vienna,
Silicone stopper 0.5 Austria.
Cable 2 Fisher D.K. 2007. Automated collection of soil-moisture data with a low-
Total 21 cost microcontroller circuit. Appl. Eng. Agric. 23:493-500.

[page 118] [Journal of Agricultural Engineering 2013; XLIV:e16]

 Technical Note

Fisher D.K, Gould P.J. 2012. Open-source hardware is a low-cost alter-
native for scientific instrumentation and research. Modern
Instrument. 1:8-20.
Fisher D.K., Kebede, H. 2010. A low-cost microcontroller-based system
to monitor crop temperature and water status. Comput. Electr.
Agric. 74:168-73.
Freescale Semiconductor. 2010. Technical Data Sheet for MPX5100. Rev
13, 05/2010. Available from: http://www.freescale.com/files/sensors/
doc/data_sheet/MPX5100.pdf?pspll=1
Grashey-Jansen S. 2012. Soil hydrology and soil properties on a partial-
ly reforested hillside in the Central Alps. J. Forest Sci. 58:363-71.
Gratton S.R., Oster J. 1992. Water quality guidelines for trees and vines.
Drought Tips 92-38. Department of Land, Air and Water Resources,
University of California, USA.
Greswell R., Ellis P., Cuthbert M., White R., Durand V. 2009. The design
and application of an inexpensive pressure monitoring system for
shallow water level measurement, tensiometry and piezometry. J.
Hydrol. 373:416-25.
Riley T.C., Endreny T.A., Halfman J.D. 2006. Monitoring soil moisture
and water table height with a low-cost data logger. Comput. Geosci.
Sanders D.C. 1997. Vegetable crop irrigation. Horticulture Information
Leaflet 33-E, North Carolina Cooperative Extension Service, North
Carolina Sate University, USA.
Testezlaf R., Larsen C.A., Yeager T.H., Zazueta, F.S. 1999. Tensiometric
monitoring of container substrate moisture status. Hort. Technol.
9:105-9.
Thalheimer M. 2003. Tensiometer modification for diminishing errors
due to the fluctuating inner water column. Soil Sci. Soc. Am. J.
67:737-9.
Thalheimer M. 2005. Zur Dynamik des Bodenwassers an einem grund-
wassernahen Obstbau-Standort im Etschtal (Südtirol, Italien) [Soil
water dynamics in an apple orchard with shallow groundwater table
in the Adige Valley (South Tyrol, Italy)]. Laimburg J. 2:50-7.
Thompson R.B., Gallardo M., Valdez R.C., Fernández R.D. 2007. Using
plant water status to define threshold values for irrigation manage-
ment of vegetable crops using soil moisture sensors. Agric. Water
Manage. 88:147-58.
Warrick A.W., Wierenga P.J., Young M.H., Musil S.A. 1998. Diurnal fluc-
tuations of tensiometric readings due to surface temperature

y
32:135-40. changes. Water Resour. Res. 34:2863-9.

l
on
e
us
al
ci
er
m
m
- co
on
N

[Journal of Agricultural Engineering 2013; XLIV:e16] [page 119]